Evaluation of livestock data using multivariate analysis techniques

The field of livestock plays a significant role in the economy of pakistan. the available univariate techniques are not competent to uncover all the features of the livestock data. thus current study was conducted to assess more features of the livestock variables measured at a livestock experiment station,rakh gulama,bhakkar,pakistan which has data on 74 holstein-friesian x sahiwal crossbred and 237 nili-ravi buffaloes for various productive and reproductive traits accumulated over 10 and 13 years respectively. the highest correlation was observed between calving interval and service period (r = 0.98),followed by calving interval and dry period (r = 0.95) and service period and dry period (r = 0.93) for different traits of buffaloes while for different traits of cattle's correlation between these variables is also high. however,the first principal component (pc1) elucidates 41.8% of the total variation for different traits of buffalo's data while it is 44.78% for different traits of cattle's data. the pc2,pc3 and pc4 for different traits of buffaloes explains 26.33%,17.67% and 12.44% of the total variation respectively while these are 24.11%,18.22% and 11.44% for various traits of cattle's data. on the basis of these results it was suggested that the variables under study can be classified into three clusters i.e. (i) dry period,calving interval,service period and overall average (ii) age at first calving and dry period,and (iii) milk yield,wet average and lacation period to mitigate the misleading results that are revealed using univariate techniques.